This thesis resulted in the development of robust chironomid-based transfer-functions for February mean air temperature and the concentration of total nitrogen (TN) in lake-water. The New Zealand transfer-functions for both variables compare favourably with chironomid-based transfer-functions for equivalent variables from elsewhere in the world, and diatom-based transfer-functions for nutrients and lake production from New Zealand. The application of the temperature and TN transfer-functions provided insight into New Zealand climate conditions during the last glacial and served as validation for the reconstructions. Chironomid-based Temperature reconstructions from lake silts preserved in the banks of Lyndon Stream indicate a maximum cooling of ca 4 ℃ between 26.6 and 24.5 ka BP, which is consistent with estimates based on beetles and plant macrofossils. A cooling of 4 ℃ is insufficient to explain the lack of canopy tree pollen in many New Zealand pollen records at this time. Other environmental parameters additional to temperature may have limited the expansion forest cover. The chironomid-based TN reconstructions infer a trend of rapidly deteriorating water-quality in a small doline in north-west Nelson, in the South Island of New Zealand following deforestation immediately surrounding the lake ca. 1970 AD. The overall trend and timing of eutrophication inferred from the chironomids was consistent with other biological proxies and actual observations of changes in lake water quality. The chironomid-based transfer-functions provide a valuable new tool for the study of longterm climate variability and improving our understanding of the response of aquatic ecosystems to long-term natural and human induced environmental change in New Zealand lakes. I have identified some possibilities for future research which should improve the performance of these transfer-functions. The improvement of the chironomid taxonomy and the expansion of the training set should be the highest priorities.

This thesis resulted in the development of robust chironomid-based transfer-functions for February mean air temperature and the concentration of total nitrogen (TN) in lake-water. The New Zealand transfer-functions for both variables compare favourably with chironomid-based transfer-functions for equivalent variables from elsewhere in the world, and diatom-based transfer-functions for nutrients and lake production from New Zealand. The application of the temperature and TN transfer-functions provided insight into New Zealand climate conditions during the last glacial and served as validation for the reconstructions. Chironomid-based Temperature reconstructions from lake silts preserved in the banks of Lyndon Stream indicate a maximum cooling of ca 4 ℃ between 26.6 and 24.5 ka BP, which is consistent with estimates based on beetles and plant macrofossils. A cooling of 4 ℃ is insufficient to explain the lack of canopy tree pollen in many New Zealand pollen records at this time. Other environmental parameters additional to temperature may have limited the expansion forest cover. The chironomid-based TN reconstructions infer a trend of rapidly deteriorating water-quality in a small doline in north-west Nelson, in the South Island of New Zealand following deforestation immediately surrounding the lake ca. 1970 AD. The overall trend and timing of eutrophication inferred from the chironomids was consistent with other biological proxies and actual observations of changes in lake water quality. The chironomid-based transfer-functions provide a valuable new tool for the study of longterm climate variability and improving our understanding of the response of aquatic ecosystems to long-term natural and human induced environmental change in New Zealand lakes. I have identified some possibilities for future research which should improve the performance of these transfer-functions. The improvement of the chironomid taxonomy and the expansion of the training set should be the highest priorities.

Aquatic eutrophication is a serious global problem, associated with phytoplankton blooms, hypoxia, and loss of species. The objective of this thesis was to advance understanding of stream and lake eutrophication within Canterbury (South Island, New Zealand). I investigated three key questions: 1) How do riparian characteristics control stream trophic state, 2) how does stream trophic state in the Canterbury region compare to stream trophic state nationally and internationally, and 3) what factors control trophic state in Te Wairewa/Lake Forsyth. I measured rates of stream community metabolism in 21 Canterbury streams over a gradient of riparian canopy cover, and conducted a literature review of national and international studies of stream metabolism. I also examined the occurrence of cyanobacterial blooms in Te Wairewa in relation to water quality and weather from 17 years of measurements, and performed series of nutrient addition assays on the lake to assess nutrient limitation. I found that riparian characteristics strongly controlled stream trophic state by shading, thereby reducing photosynthetic productivity. This overwhelmed the effects of high nitrate concentrations, which increased primary production. Compared to national and international rates of stream metabolism, Canterbury streams were strongly heterotrophic, with low rates of autotrophic production. Catchment streams draining into Te Wairewa were unlikely to be the main source of nutrients supporting large cyanobacterial blooms. Instead, internal lake nutrient loading mechanisms associated with calm weather were likely to supply blooms. My results emphasize the importance of light limitation, nitrogen and heterotrophy in controlling stream trophic state, and nutrient supply and weather in controlling lake trophic state.